A provider is the vendor that serves a model - for example OpenAI, Anthropic, Fireworks AI, Together AI, or Infercom. Each provider has its own API format, authentication, and pricing. An LLM gateway normalizes those differences behind one API and one bill.
Why it matters
Providers differ in API format, authentication, model lineup, and price. Without a gateway, supporting several providers means several integrations, keys, and bills to maintain.
How it works
A gateway wraps each provider behind an adapter that translates your one request into the provider's format and normalizes the response. Infere routes across OpenAI, Anthropic, Fireworks AI, Together AI, and Infercom through a single API.